1. Field of the Invention
This invention relates to an exposure method for use in photolithography and a mask for use in photolithography. More particularly, the invention relates to a phase-shift mask and an exposure method using the phase-shift mask.
2. Description of the Related Art
To form patterns of semiconductor elements, a photolithography technique is commonly employed. A pattern of a mask is transferred to a photosensitive resin layer provided on a semiconductor substrate by the photolithography technique. The photosensitive resin is also known as xe2x80x9cresist.xe2x80x9d A resist is classified into two type, i.e., negative type and positive type. The negative-type resist is of the type; any part of which that has been exposed to the light applied through a mask will remain on the semiconductor substrate. The positive-type resist is of the type; any part of which that has been exposed to the light applied through a mask will be removed from the semiconductor substrate.
In recent years, it has been demanded that an image be formed on a resist layer in higher resolution to provide fine patterns of semiconductor elements. Fine semiconductor element patterns increase integration density of a semiconductor integrated circuit.
To enhance the resolution of an image formed on a resist, a phase shift exposure method is proposed in 1982. In the phase shift exposure method, the phase difference between light beams applied is utilized to improve the resolution of the image focused on a resist layer. The principle of the phase shift exposure will be described, with reference to FIGS. 1A to 1D and FIGS. 2A to 2D.
In the ordinary exposure, the light applied perpendicularly to a mask 106 passes through the transparent regions 150 and 151 of the mask as illustrated in FIG. 1A. Chromium mask patterns 121 are provided on the mask 106. The mask 106 has transparent regions 150 and 151. The light beams passing through the transparent regions 150 and 151 have the same phase. The light beams emanate from the transparent regions 150 and 151 and pass through the projection lens of a reducing projection exposure apparatus. The two beams are then focused on the surface of a resist layer, which is on an image-forming surface.
The distance between the transparent regions 150 and 151 cannot be reduced to an infinitesimal value, for the following reason. If the distance is extremely short, the two beams passing the regions 150 and 151 overlaps at the image-forming surface as indicated by the broken lines in FIG. 1C. The light beams, which have a same phase, intensify each other at the image-forming surface. As a result, the light-intensity distribution on the surface of the resist has one peak as illustrated in solid line in FIG. 1D. Consequently, the chromium mask patterns 121 are not correctly transferred to the resist layer. Thus, the interval between the transparent regions 150 and 151 cannot be decreased over a certain limit. The limit R of resolution for any image formed on a resist is given as follows:
R=K1xc3x97xcex/NAxe2x80x83xe2x80x83(1) 
where K1 is the constant that depends on the properties of the photosensitive resin, xcex is the wavelength of the light applied to the mask 106, and NA is the numerical aperture of the projection lens that is incorporated in the reducing projection exposure apparatus. Here, the limit R is known as xe2x80x9cReyleigh resolutionxe2x80x9d.
In the phase shift exposure, light is applied to a resist layer through a phase shift mask 107 as is illustrated in FIG. 2A. The phase shift mask 107 has transparent regions 152 and 153. The region 153 is provided with a phase shifter 120, while the region 152 has no phase shifters. The light beam passing through the transparent region 153 is delayed as it passes through the phase shifter 120. Hence, the light beam passing through the transparent region 153 differs in phase from the light beam passing through the transparent region 152. The thickness D that the phase shifter should have to impart a phase difference of 180xc2x0 to the light beams is given as follows:
D=xcex/{2xc3x97(nxe2x88x921)}xe2x80x83xe2x80x83(2) 
where xcex is the wavelength of the light applied to the phase shift mask 107, and n is the refractive index of the phase shifter 120. If the two light beams emanating from the transparent region 152 and the transparent region 153, respectively, have a phase difference of 180xc2x0, their parts overlapping at the image-forming surface will cancel out each other. As a result, as shown in FIG. 2C, the intensity of light is nil at one part of the surface of the resist layer. It follows that the light-intensity distribution on the resist has two peaks as shown in FIG. 2D. The chromium patterns 121 can therefore be transferred to the resist with high accuracy. Thus, the use of the phase shift mask 107 can enhance the resolution of an image focused on the surface of a resist.
Also, the phase shift exposure technique can increase the depth of focus (DOF). The term xe2x80x9cdepth of focusxe2x80x9d means the range of distance over which the focus may be displaced without causing troubles. The reason is discussed comparing the ordinary exposure technique and the phase exposure technique in the following.
In the ordinary exposure using no phase shifters, the more the image-forming surface deviates from the focal plane, the more the two beam emanating from the transparent regions 150 and 151 overlap each other at the image-forming surface. This means that the resolution will sharply decrease if the image-forming surface of the resist deviates from the focal plane.
In the phase shift exposure, the two adjacent beams emanating from the transparent regions 152 and 153 have a phase difference of 180xc2x0. Their overlapping parts cancel out each other at the image-forming surface of the resist layer. The intensity of light is therefore zero at one part of the image-forming surface. Hence, even if defocusing occurs, that is, even if the focus deviates from the image-forming surface, the dimensional precision of the pattern, transferred to the resist, will be hardly influenced. Thus, the depth of focus can be increased in the phase shift exposure.
The phase shift exposure technique, however, cannot successfully apply to two-dimensional random patterns. The layout pattern of a semiconductor integrated circuit includes regular patterns and random patterns. Each regular pattern extends in one direction only, whereas each random pattern randomly extends first in one direction, and then in another direction. Here, examples of regular patterns are the bit lines and word lines of a DRAM (Dynamic Random Access Memory). Examples of random patterns are the wires of logic circuits. The phase shift masks are designed in accordance with the basic rule that a phase difference of 180xc2x0 is imparted to two beams that have passed through two adjacent transparent regions. This basic rule can be easily applied to the regular patterns, but not to two-dimensional random patterns.
FIG. 3A is a plan view of a phase shift mask 108 that may be used to form two-dimensional random patterns by means of the conventional phase shift exposure. The phase-shift mask 108 is designed to transfer a pattern on a positive-type resist. The mask 108 has a shield region 111, a transparent region 113 and a transparent region 114. The shield region 111 is identical in shape to the pattern that is to be transferred to a resist. Phase shifters 120 are provided on the transparent region 113. The beam passing through the transparent region 113 is out of phase with respect to the beam passing through the transparent region 114. In other words, the phase of the beam differs by 180xc2x0 from that of the beam passing through the transparent region 114. FIG. 3C is a sectional view of the Levenson-type mask 108, taken along line Cxe2x80x94C in FIG. 3A. As FIG. 3C shows, the mask 108 is composed of a glass substrate 122. A chromium film 121 is provided on the shield region 111, and a phase shifter 120 is provided on the transparent region 113. No phase shifters are provided on the transparent region 114.
Light is applied to the positive-type resist through the phase-shift mask 108. Light-exposed parts of the positive-type resist are developed and resist patterns 117 are formed as shown in FIG. 3B. The shield region 111 shields the part 115 of the positive-type resist from the light. Thus, the part 115 of the resist, which opposes the shield region 111, is developed as is intended.
In addition, the part 116 of the resist layer, which opposes the boundaries between the transparent regions 113 and 114, is developed. The beams passing through the transparent region 113 (having a phase shifter) and the transparent region 114 (having no phase shifters) have the opposite phases. The intensity of light is therefore almost nil at the boundary between the transparent regions 113 and 114. The part 116 of the resist, which opposes the boundary between the transparent regions 113 and 114, is also developed. That is, not only the part 115 that should be developed, but also the parts 116 which should not be developed is developed.
With the conventional phase shift exposure it is difficult to prevent the part 116 of the resist layer from being developed. Hence, the conventional phase-shift mask 108 cannot be used to transfer two-dimensional random patterns to a positive-type resist.
It will be described now how two-dimensional random patterns are transferred to a negative-type resist by means of the conventional phase shift exposure technique. FIG. 4 is a plan view of a phase-shift mask 109 that is used to transfer two-dimensional random patterns to a negative-type resist. As shown in FIG. 4, the phase-shift mask 109 has a shield region 111xe2x80x2, a transparent region 113xe2x80x2 and a transparent region 114xe2x80x2. A phase shifter is provided on the shield region 111xe2x80x2. The transparent region 113 has phase shifters while the transparent region 114xe2x80x2 has no phase shifters. The phase shifter imparts a phase different of 180xc2x0 to the beam that has passed through the transparent region 113xe2x80x2, with respect to the beam that has passed through the transparent region 114xe2x80x2.
The transparent region 113xe2x80x2 is an auxiliary pattern for enhancing the resolution of the negative-type resist pattern that will be formed at a position corresponding to the transparent region 114xe2x80x2. Nonetheless, the transparent region 113xe2x80x2 is required to have a width equal to or less than the value equivalent to the resolution limit. It is difficult to form transparent regions having such a small width at high reliability. Hence, the conventional phase shift exposure technique cannot process resists to form two-dimensional random patterns on negative-type resists.
As described above, the conventional phase shift exposure cannot be applied to transfer two-dimensional random patterns on positive-type resists or negative-type resists.
It is desired that images be formed on resists at resolution high enough to form two-dimensional random patterns.
It is also desired that two-dimensional random patterns of high precision be transferred to resists by means of phase shift exposure technique.
An object of the present invention is to form high-resolution images on resists in the process of forming two-dimensional random patterns.
Another object of the invention is to transfer two-dimensional random patterns of high precision to resists by means of phase shift exposure technique.
In order to achieve an aspect of the present invention, a method of forming a photoresist pattern by a photolithography technique is composed of:
providing a photoresist layer;
exposing the photoresist layer to a first pattern-defining light using a first mask; and
exposing the photoresist layer to a second pattern-defining light using a second mask. The first mask includes a shielding region shielding the first pattern-defining light. The second mask includes a phase-shifting region having a phase shifter edge and a non-phase-shifting region adjacent to the phase-shifting region on the phase shifter edge. A first light portion of the second pattern-defining light passes through the phase-shifting region. A second light portion of the second pattern-defining light passes through the non-phase-shifting region. A first phase of the first light portion differs from a second phase of the second light portion. The first and second masks are aligned such that the phase shifter edge overlaps on the shielding region.
The shield region may include a line resist shielding portion to form a line resist pattern extending to a first direction. The line shielding portion has a centerline extending to the direction. In this case, it is desirable that the phase shifter edge substantially overlaps on the centerline when the first and second masks are aligned.
The phase shifter edge may be composed of first and second phase shifter edges parallel to each other and extending to the first direction. In this case, a distance between the first and second phase shifter edges is desirably larger than a width of the line shielding portion in a second direction perpendicular to the first direction.
The phase-shifting region may be provided with a phase-shifter layer. In this case, a thickness of the phase-shifter layer is desirably determined such that a phase difference between the first phase and the second phase ranges from 175 to 185xc2x0.
Also, the first pattern-defining light has a first intensity and the second pattern-defining light has a second intensity. In this case, the second intensity is desirably larger than the first intensity.
In order to achieve another aspect of the present invention, A mask set is composed of a first mask and second mask. The first mask includes a shielding region shielding a pattern-defining light exposed to the first mask. The second mask includes a phase-shifting region having a phase shifter edge and a non-phase-shifting region adjacent to the phase-shifting region in the phase shifter edge. A first phase of a first light passing through the phase-shifting region differs from a second phase of a second light passing through the non-phase-shifting region. The phase shifter edge overlaps on the shield region when the first and second masks are aligned.